Food and Nutrient Intake Response To Food Prices and Govt Programs

DECEMBER 2022
DISCUSSION PAPER SERIES NO. 2022-44
Food and Nutrient Intake Response to Food

Prices and Government Programs: Implications
for the Recent Economic Shocks
Roehlano M. Briones
The PIDS Discussion Paper Series constitutes studies that are preliminary and subject to further revisions. They are being circulated in a limited number of copies only for
purposes of soliciting comments and suggestions for further refinements. The studies under the Series are unedited and unreviewed. The views and opinions expressed are those
of the author(s) and do not necessarily reflect those of the Institute. Not for quotation without permission from the author(s) and the Institute.
CONTACT US:
RESEARCH INFORMATION DEPARTMENT
Philippine Institute for Development Studies
18th Floor, Three Cyberpod Centris - North Tower publications@pids.gov.ph

https://www.pids.gov.ph
EDSA corner Quezon Avenue, Quezon City, Philippines (+632) 8877-4000
Food and Nutrient Intake Response to Food Prices
and Government Programs: Implications for the Recent
Economic Shocks
Roehlano M. Briones
PHILIPPINE INSTITUTE FOR DEVELOPMENT STUDIES
December 2022
Abstract
The study aims to estimate the short-term impact of the recent rounds of food price inflation
and the COVID19 crisis by estimating a household food demand system, converted to energy
and nutrient intakes. It demonstrates that recent food and nutrition surveys conducted by the
Philippine government are a valuable source of information about household behavior and the
impact of economic shocks. It offers a novel methodology for incorporating computing
selection effects in determinants in estimating price and income elasticities. The study was
nevertheless able to determine that the COVID19 social protection programs played a
significant role in preventing further deterioration in nutrient intakes and worsening of
malnutrition. Notwithstanding rapid economic growth, the recent inflation episodes pose a
major threat to nutrient intakes and nutrition security. Income policies in the form of targeted
cash transfers are an important, albeit expensive way to counter adverse nutrition impacts of
economic contraction. Ameliorating the impact of price increases during inflation episodes
should assume priority in policy research and response.
Keywords: food demand, economic contraction, price inflation, price and income elasticity,
nutrient intake, food policy
i
Table of Contents
1. Introduction .................................................................................................................... 1
2. Recent economic trends ............................................................................................... 2

2.1 Overview ................................................................................................................... 2
2.2 Output and employment ........................................................................................... 2
2.3 Poverty ..................................................................................................................... 4
3. Recent trends in food consumption and nutrition ..................................................... 5

3.1 Food consumption and expenditure ......................................................................... 5
3.2 Nutritional status ....................................................................................................... 9
3.3 Nutrient intake ........................................................................................................ 11
3.4 Government programs ............................................................................................ 16
4. Review of literature and methodology ....................................................................... 17

4.1 Conceptual framework ............................................................................................ 17
4.2 Studies focused on nutrient intake ......................................................................... 18
4.3 Consumption studies .............................................................................................. 19
4.4 Role of government programs ................................................................................ 20
5. Empirical strategy for household demand ................................................................ 20

5.1 Data source ............................................................................................................ 20
5.2 Case of purchased food groups ............................................................................. 22
6. Results of demand estimation .................................................................................... 26

6.1 The selection model for household food demand ................................................... 26
6.2 Food demand estimation ........................................................................................ 27
6.3 Elasticities of energy and nutrient intake ................................................................ 30
7. Simulated impacts of recent economic shocks: household level .......................... 32

7.1 Set-up of scenarios ................................................................................................. 32
7.2 Results .................................................................................................................... 32
8. Conclusion ................................................................................................................... 35
8.1 Contributions and limitations of the study ............................................................... 35
8.2 Implications for policy ............................................................................................. 36
9. Bibliography ................................................................................................................. 37
10. Appendix Tables ...................................................................................................... 40

ii
List of Tables
Table 1: Year-on-year quarterly growth of GDP by expenditure component, 2018 – 2022 (%)
................................................................................................................................................ 3
Table 2: Annual inflation rate, 2019 – 2022 (%) ...................................................................... 4
Table 3: Per capita consumption of selected items, Philippines ............................................. 6
Table 4: Per capita consumption of food items, 1978 – 2019 ................................................. 6
Table 5: Mean one-day household food consumption by food group and wealth quintile,
Philippines, 2018..................................................................................................................... 7
Table 6: Mean food cost of households, by major food group and wealth quintile, Philippines,
2018 - 19 ................................................................................................................................. 8
Table 7: Expenditure shares of the major food groups, .......................................................... 9
Table 8: Share of households whose per capita energy and nutrient intake meet REI and
EAR, Philippines, 2018-19 (%) ............................................................................................. 12
Table 9: Contribution to energy and nutrient intake of households, by food group,
Philippines, 2018-19 (%) ....................................................................................................... 14
Table 10: Share of household intake of energy and selected nutrients from cereals and fish,
meat and poultry, by wealth quintile (%) ............................................................................... 15
Table 11: Share of individuals meeting REI and EAR, by age group, Philippines, 2018-19
(%) ........................................................................................................................................ 15
Table 12: Change in share of individuals meeting REI and EAR by age group, Philippines,
2015 vs 2018-19 (percentage points) ................................................................................... 16
Table 13: Share of households and individuals participating in selected government
programs, Philippines, 2018-19 (%) ...................................................................................... 17
Table 14: Mean values of selected variables, by food group, FCS ....................................... 22
Table 15: Coefficients of price and expenditure variables, probit regressions ...................... 26
Table 16: Coefficients of price and expenditure variables, probit regressions ...................... 27
Table 17: Elasticity of household consumption with respect to food expenditure and income,
with and without selection effects ......................................................................................... 28
Table 18: Price elasticities of household demand, with selection effects ............................. 28
Table 19: Price elasticities of household demand, without selection effects ........................ 29
Table 20: Intake elasticities with respect to price, energy and selected nutrients ................ 31
Table 21: Coverage and benefit estimates for Bayanihan social protection programs, 2020 32
Table 22: Changes in intake of energy and selected nutrients, COVID19 2020 scenario,
household and selected members (%) ................................................................................. 33
Table 23: Changes in intake of energy and selected nutrients, COVID19 with Bayanihan,
Table 24: Changes in intake of energy and selected nutrients, Inflation with growth scenario,
iii
List of Figures
Figure 1: Quarterly rate of unemployment and visible underemployment, 2019 – 2021 (%) .. 3
Figure 2: Poverty incidence of families and of population, Philippines, 2015 – 2021 (%) ....... 5
Figure 3: Stunting prevalence among children under-5, by degree of stunting and wealth
quintile of household (%) ...................................................................................................... 10
Figure 4: Prevalence of anemia and Vitamin A deficiency among children, pregnant, and
lactating women, Philippines, 2018-19 (%) ........................................................................... 11
Figure 5: Prevalence of iron and Vitamin A deficiency among children under-5, by wealth
quintile, Philippines, 2018-19 (%) ......................................................................................... 11
Figure 6: Daily Cost of Recommended Diet and Expenditure per Adult Male Equivalent, 2015
.............................................................................................................................................. 13
Figure 7: Cost per 1,000 kcal by food category, Philippines and Southeast Asia, 2011 ....... 13
Figure 8: Individual-to-household consumption elasticities, by household member type ..... 30
Figure 9: Intake elasticities with respect to income, energy and selected nutrients.............. 31
iv
Food and Nutrient Intake Response to Food Prices and Government Programs:
Implications for the Recent Economic Shocks
Roehlano M. Briones*
1. Introduction
Even before the Coronavirus 2019 (COVID19) pandemic, the nutritional status of Filipino
children was already serious concern, with the stunting rate at 28.8 percent in 2019 (DOST-
FNRI, 2019). In March 2020, the COVID19 pandemic reached the Philippines. As a result of
the COVID19 community quarantines and worldwide economic contraction, Gross Domestic
Product (GDP) underwent a unprecedented decline in 2020. Though the economy recovered
somewhat in 2021, GDP in 2nd quarter 2022 is still lower by 5.6 percent compared to 4th
quarter 2019. The economic contraction has likely led to dramatic increases in poverty, though
the Social Amelioration Package (SAP) of the government likely mitigated the decline (Reyes,
2021). Following the output and employment contraction in 2020 was rising inflation in 2021-
22, as world prices of crude oil, fertilizer, and food reached historic peaks. These economic
shocks are likely to have had adverse impact on food consumption and nutrient intake at the
household level, leading to worsening nutritional status. This will push the Sustainable
Development Goal (SDG) of ending hunger and malnutrition by 2030 further out of reach.
There is however little information about the size of household level consumption impacts of
these economic shocks. This study aims to establish an empirical relationship between
economic adjustments in the form of prices, income and expenditure, and government
programs, and household and individual food consumption and nutrient intake. In doing so, it
will provide a valuable input in the design of short- and long-term programs for addressing
household food insecurity and malnutrition. The key evaluation questions for this study are as
follows:
• By how much did household food consumption and nutrient intake decline in response
to recent economic shocks?
• By how much does individual level consumption and nutrient intake vary given changes
in household consumption?
• What can we infer about the design of government programs in terms of ensuring food
security and proper nutrition in the event of economic shocks?
The study aims to estimate the short-term impact of the recent rounds of food price inflation
and the COVID19 crisis. In doing so it develops a policy tool that may also serve to evaluate
nutrition impacts of long term economic change, and its implication for achieving SDG 2 by
2030, namely ending hunger and malnutrition. Specifically, the study aims to:
1. Examine patterns of response of household and individual food consumption and nutrient
intake to changes in food prices, household income and expenditure, and public programs;
2. Estimate elasticities of consumption and nutrient intake with respect to price and
expenditure and possibly income;
3. Estimate transmission elasticities from household price and expenditure to individual food
consumption and nutrient intake;
1
4. Based on these patterns and responses, to estimate the impact of recent bouts of inflation
and economic contraction in 2020 - 2022, as well as accompanying safety nets, on
household food consumption and nutrient intake;
5. Draw policy implications about the design of government programs in terms of ensuring
food security and proper nutrition in the event of economic shocks.
2. Recent economic trends
2.1 Overview
This section reviews economic and household welfare indicators from secondary official
sources, focusing on the most recent period. Philippine Statistics Authority (PSA) official
statistics provide the economic indicators, as well as trends in poverty and per capita
consumption in selected years. Per capita consumption is obtained by recall method, in raw,
as-purchased form, over the past week.
Another estimate of per capita consumption uses the one-day intake as measured by direct
weighing under the National Nutrition Survey (NNS) conducted by the Food and Nutrition
Research Institute (DOST-FNRI) of the Department of Science and Technology (DOST). Data
on household consumption measures the quantity of food reckoned in raw, as-purchased form,
but accounts for household (plate) wastage, hence the resulting figure pertains to consumption
as quantity of food actually eaten, expressed in raw form. In 2018-19, DOST-FNRI conducted
the Expanded NNS, which covers information on: socioeconomic status; food consumption
and nutrient intake; anthropometric measures; biomedical measures; clinical and health
practices; maternal health and nutrition; infant and young child feeding practices; food security;
and participation in government programs.
2.2 Output and employment
GDP underwent severe contraction in 2020, led by a fall in investment, though 2021 has seen a
strong recovery
Quarterly GDP growth (year-on-year) was still a decent 6.1% average for 2019, but a
contraction already began in first quarter of 2020, deepening to an unprecedented level in the
second quarter. This continued over the succeeding quarters, reversing only the first quarter of
2021. Growth has remained consistently high over the succeeding quarters. By component,
during the contraction phase all components also fell sharply except for Government
consumption; the worst performing were investment and exports. In the recovery phase, all the
components also recovered, with the biggest recovery for investment.
The COVID-19 pandemic was likewise associated with a sharp rise in unemployment and
underemployment, which had recovered to pre-pandemic levels by late-2021.
Quarterly unemployment rate held steady at around 5% over 2019 and first quarter of 2020.
The visible underemployment rate, which is the share of workers who work under forty hours
a week and seek more hours of work per week, had also been fairly steady at around 8-9
percent. In the second quarter there was a sharp spike in both measures, with rate of
unemployment even exceeding that of visible underemployment, although the decline had
already begun even in the third quarter. By 2021, unemployment remained elevated above 2019
levels, whereas visible underemployment again began to rise in early 2021, before dipping even
below 2019 levels by late 2021.
2
Table 1: Year-on-year quarterly growth of GDP by expenditure component,
2018 – 2022 (%)
2019 2020 2021 2022
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4
Household consumption 5.9 0.2 -15.3 -9.2 -7.3 10.0 8.6 13.8 7.8 7.4
Government consumption 9.3 7.0 21.8 5.8 5.0 3.6 11.1 20.8 14.2 20.5
Gross capital formation 3.7 -12.3 -51.8 -38.8 -31.7 20.4 20.5 9.1 7.7 7.4
Exports 2.6 -3.6 -33.2 -15.3 -10.7 10.4 4.3 12.7 14.3 14.5
Imports 2.5 -7.2 -37.2 -20.6 -20.6 15.4 13.6 7.0 7.8 7.8
GDP 6.1 -0.7 -16.9 -11.6 -8.2 8.2 7.4 7.1 7.5 9.3
Note: 2022 pertains to the average of quarters 1 and 2.
Source: PSA (2022).
Figure 1: Quarterly rate of unemployment and visible underemployment, 2019 – 2021 (%)
Source: PSA (2022).
The Bayanihan programs of government provided substantial social protection to households in

2022.
The Bayanihan Heal as One Act (Bayanihan 1) was passed by late March, 2020, within a couple
of weeks of the declaration of the State of National Calamity owing to COVID19. The law
authorized a cash transfer program known as SAP, provided to beneficiaries of the Pantawid
Pamilyang Pilipino Program (4Ps) of the government, the on-going social protection scheme
aimed at the poor. In September the same year, the Bayanihan II was passed, continuing social
protection schemes of Bayanihan 1. It also expanding labor market measures through cash for
work, such as the Tulong Panghanapbuhay sa Ating Disadvantaged/Displaced Workers
(TUPAD and access to capital at concessional rates aimed at micro, small, and medium
enterprises. These social protection measures made up the bulk of the fiscal response to
COVID19, amounting to about 2.5 percent of GDP (Cho and Johnson, 2022).
3
Several key food items have undergone severe price inflation since 2021.
Economic contraction in 2020 compounded an on-going bout with inflation starting in 2022.
Previously in 2018, inflation had been driven by rice prices at 7.3 percent, whereas the overall
inflation rate was 4.3 percent. In 2019, rice prices actually declined owing to Republic Act
(RA) 11203, the Act liberalizing the rice industry. Prices remained generally stable through to
2021 (Table 2). Inflation started to become a problem across-the-board in 2022, with the overall
rate (in annual terms reaching 7.7%, breaching by far the Bangko Sentral ng Pilipinas (BSP)
ceiling of 4 percent. Fats & oil hit 20 percent, followed by Vegetables at 15 percent, Meat
reached 12 percent, and wheat-based products at 10 percent, followed closely by Fish and
Dairy. Not shown in the Table is sugar and confectionary products, whose price index rose by
an astonishing 33 percent from January to October 2022. Overall the food inflation rate was 10
percent. The main drivers were world prices of crude oil and cereals; even rice reversed to
positive growth in 2022 (though at a mild 2.5 percent) as world fertilizer prices began to exert
pressure on rice prices.
Table 2: Annual inflation rate, 2019 – 2022 (%)

2019 2021 2022
Rice -4.2 -0.7 2.5
Flour, bread, bakery products 3.5 2.0 9.8
Fish 3.9 6.4 9.4
Meat 3.8 13.4 11.5
Fruits 5.2 -0.1 4.9
Vegetables 2.7 4.8 16.0
Dairy 2.4 1.1 8.7
Fats & oil 1.6 5.6 20.4
Other food 4.5 1.3 8.1
Food 1.5 4.5 9.8
All items 2.4 3.9 7.7
Note: 2022 figures are computed as the average of montly year-on-year inflation rates for January to October.
Source: PSA (2022).
2.3 Poverty
After a sharp decline in 2015 – 2018, poverty increased in 2018-21, though the 2021 level is
still below that of 2015.
One may easily guess that household welfare was negatively impacted by these economic
shocks. This is confirmed by poverty estimates for 2021 (Figure 2).
4
Figure 2: Poverty incidence of families and of population, Philippines, 2015 – 2021 (%)
Source: PSA (2022).
For the population, the incidence of poverty was 18.1 percent, up from the historic low of 16.7
percent in 2018. Likewise, the incidence of family poverty was 13.2 percent in 2018, up from
12.1 percent in 2018. Fortunately the reversal was not sharp enough to reverse all the poverty
gains made from 2015, probably because of the strong recovery of employment and economic
growth in 2021.
3. Recent trends in food consumption and nutrition
3.1 Food consumption and expenditure
Before the recent economic shocks, per capita consumption of several key food items had been
declining.
PSA survey data on per capita consumption is available from 1999 onward for selected years
(Table 3). Up to 2012 the PSA (then through the former Bureau of Agricultural Statistics)
conducted a Survey of Food Demand, while in 2018, PSA used the Family Income and
Expenditure Survey (FIES) to generate measures of per capita consumption. Per capita
consumption had been rising up to 2012 for Rice, Pandesal, Eggplant, Ampalaya, Chicken,
Milkfish, Tilapia, and Roundscad. However from 2012 to 2018, PSA recorded a decline in per
capita consumption, though some caution must be placed on interpreting these estimates given
the difference in data source. Nonetheless there were sharp reductions in Rice (9.6 percent),
Pork (32.6 percent), Chicken (14.8 percent), and Roundscad (31.6 percent).
5
Table 3: Per capita consumption of selected items, Philippines
1999-2000 2008-2009 2012 2018
in kg/year
Rice 105.8 119.1 114.3 103.3
Corn 10.9 7.1 10.3 9.1
Pandesal . 4.2 6.3 .
Camote 7.4 4.1 4.3 .
Cassava 6.9 3.1 2.8 .
Eggplant 3.9 4.2 4.1 .
Ampalaya 1.1 2.5 2.4 .
Tomato 3.2 2.6 3.2 .
Banana 22.4 16.7 . .
Mango 3.3 3.2 3.5 .
Pork 11.2 9.2 9.5 6.4
Beef 2.8 1.2 0.9 .
Chicken 7.4 7.9 8.1 6.9
Milkfish 2.9 4.2 3.7 3.0
Tilapia . 4.7 4.8 3.6
Roundscad . 5.4 5.7 3.9
Sources: PSA (2022).
In the 2010s, per capita consumption of most major food groups has been falling, except
vegetables.
The remaining tables for consumption at the household level are derived from NNS, where the
series starts in 1978 up to 2018-2019 (DOST-FNRI, 2022a and 2022b). NNS were conducted
every five years except in 1982, and 2018-19. For the major food groups per capita
consumption had been falling over the thirty-year period from 1978 to 2018-19, except for
meat, where per capita consumption more than doubled over the period (Table 4). However in
the 2010s, even meat consumption (together with the rest of the food groups) also suffered a
decline in per capita consumption.
Table 4: Per capita consumption of food items, 1978 – 2019

1978 1982 1987 1993 2003 2008 2013 2015 2018-19
In gm/day
Cereals & cereal products 367 356 345 340 364 361 346 358 315
Rice & rice products 308 304 303 282 303 317 299 308 276
Fish & fish products 102 113 111 99 104 110 109 101 94
Meat & meat products 23 32 37 34 61 58 65 61 58
Vegetables 145 130 111 106 111 110 114 123 126
Fruits 104 102 107 77 54 54 41 37 34
Sources: DOST-DOST-FNRI (2022a).
6
The following discussion focuses on data from the 2018-19 NNS. Disaggregated consumption
estimates are available at the household level, while nutrient intake and nutritional status are
available at the individual level. Mean household size was 4.2 members. The NNS also collects
data on household assets, which are combined using principal component analysis into a wealth
index to sort households into quintiles, referred to as Poorest, Poor, Middle, Rich, and Richest.
Increasing wealth is associated with lower consumption of cereals and vegetables, and rising
consumption of animal products, fats & oils, and fruit.
Estimates of mean one-day food consumption at the household level for 2018-19 are shown in
Table 5. Average daily intake for a family of 4.5 members is about 3 kg of food, of which 1.2
kg is Cereals & cereal products, of which 88 percent is rice by weight. This is followed by Fish
and fish products at 339 gms per day; on annual basis per capita this amounts to about 29.5 kg
of fish & fish products, making this food group by far the most significant source of animal
protein. The next most significant is Meat & meat products, followed by Milk & milk products,
then Poultry. In per capita terms, Vegetable and Fruit consumption add up to 133 gm per day,
a stunning 67 percent below the World Health Organisation (WHO) dietary recommendation
of 400 gms of fruits and vegetables per day (preferrably in five different servings).
Household consumption of Rice, Sugars & syrups, and Vegetables is highest for the Poorest
quintile, and least for the Richest quintile, implying that (holding prices constant), these food
groups are inferior goods (increasing income is associated with lower quantity consumed).
Meanwhile household consumption of Fish & fish products, Meat & meat products, Poultry,
Eggs, Milk & milk products, and Fruits are increasing with wealth quintile; the biggest
difference between the household consumption of the Richest and of Poorest is largest for Meat
& products (320 percent), followed by Poultry (294 percent), and Milk & milk products
(238 percent).
Table 5: Mean one-day household food consumption by food group and wealth quintile,
Philippines, 2018
All Poorest Poor Middle Rich Richest
(in gms per day)
Cereals & cereal products 1,213 1,269 1,254 1,230 1,172 1,105
Rice & rice products 1,064 1,068 1,094 1,098 1,053 985
Starchy roots & tubers 38 48 28 30 39 48
Sugars & syrups 32 34 33 32 31 30
Fats & oils 53 42 54 55 55 68
Fish & fish products 339 304 330 344 351 377
Meat & meat products 214 87 151 227 286 366
Poultry 113 50 77 111 158 197
Eggs 82 59 81 89 90 95
Milk & milk products 188 91 139 196 248 308
Vegetables 468 503 467 453 436 481
Fruits 122 112 89 119 128 178
Others 159 135 147 169 167 176
Total 3,021 2,734 2,850 3,055 3,161 3,429
Sources: DOST-DOST-FNRI (2022a).
7
As household wealth increases, food expenditure rises for all the food groups.
The NNS collects data on household spending on food, called household food cost; items
consumed but not purchased, e.g. produced at home, or received as gifts, are imputed a price
at the prevailing local market price. For all the major food groups, household spending
increases by wealth quintile of household (Table 6). One-day food spending of Richest
households is about Php 366, more than double that of Poorest households (Php 181) and 44
percent higher than average (Php 254). The fact that spending increases for inferior goods
implies increasing unit value (in Php per gm), probably capturing increasing quality of Rice &
rice products, Sugars & syrups, and Vegetables.
Table 6: Mean food cost of households, by major food group and wealth quintile,
Philippines, 2018 - 19
In Php per day
Cereals & cereal products 71.5 66.9 68.8 69.7 72.5 82.4
Rice & rice products 50.4 49.3 50.1 50.1 50.4 52.4
Starchy roots & tubers 2.2 1.6 1.2 1.9 2.8 4.2
Sugars & syrups 4.0 2.9 3.8 3.9 4.5 5.5
Fats & oils 5.3 3.9 4.6 5.2 6.2 7.3
Fish & fish products 42.8 32.3 28.0 43.2 47.1 58.1
Meat & meat products 37.1 14.3 24.4 37.7 49.8 62.4
Poultry 18.1 7.8 12.2 17.1 25.3 33.0
Eggs 10.4 7.4 10.6 10.7 11.2 13.0
Milk 11.1 4.8 7.3 10.0 14.5 22.1
Vegetables 26.6 22.5 24.0 25.8 28.4 34.4
Fruits 5.6 2.9 3.7 5.0 6.5 11.3
Others 18.9 14.1 27.2 19.9 17.2 32.1
Total 253.6 181.3 215.6 250.1 286.0 365.8
Sources: DOST-FNRI (2022a).
Food consumption is concentrated in animal proteins and cereals; increasing wealth is

associated with decreasing expenditure shares of inferior goods.
The average household devotes nearly half of food expenditure on animal proteins (47.1
percent); another 28 percent is allocated to cereals, of which rice alone is already about 20
percent (Table 7). Vegetables accounts for about a tenth of the food budget. As wealth quintile
increases, expenditure share of cereals and vegetables tend to decline; the Poorest quintile
spends about 27 percent of its food budget for rice, and 12.4 percent for vegetables, compared
with only 14 percent and 9 percent, respectively, for the richest quintile.
8
Table 7: Expenditure shares of the major food groups,
Cereals & cereal products 28.2 36.9 31.9 27.9 25.4 22.5
Rice & rice products 19.9 27.2 23.2 20.0 17.6 14.3
Starchy roots & tubers 0.9 0.9 0.6 0.7 1.0 1.1
Sugars & syrups 1.6 1.6 1.7 1.5 1.6 1.5
Fats & oils 2.1 2.1 2.1 2.1 2.2 2.0
Fish & fish products 16.9 17.8 13.0 17.3 16.5 15.9
Meat & meat products 14.6 7.9 11.3 15.1 17.4 17.1
Poultry 7.1 4.3 5.6 6.8 8.8 9.0
Eggs 4.1 4.1 4.9 4.3 3.9 3.6
Milk 4.4 2.6 3.4 4.0 5.1 6.0
Vegetables 10.5 12.4 11.1 10.3 9.9 9.4
Fruits 2.2 1.6 1.7 2.0 2.3 3.1
Others 7.5 7.7 12.6 8.0 6.0 8.8
Total 100.0 100.0 100.0 100.0 100.0 100.0
Sources: DOST-FNRI (2022a).
3.2 Nutritional status
The Expanded NNS provides information on nutritional status using Anthropometric Survey
and Biochemical Survey. Among the measurements made, the focus here are on the most
critical malnutrition indicators, namely stunting prevalence among children under-5, and
prevalence of iron and Vitamin A deficiency.
Prevalence of stunting among children under-5 remains seriously high, reaching alarming levels
among the Poorest quintile.
Stunting among children under-5 (referring to children 0 – 59 months old whose height-for-
age is two standard deviations below the median of WHO Child Growth Standards) represents
chronic undernutrition in the critical early period of a person’s growth. It is predictive of
economic outcomes related to cognitive development and health, such as test scords, education,
and wages (Horton and Hoddinott, 2014). In the Philippines, childhood stunting prevalence
was 44.7 percent in 1989, falling gradually to 33.1 percent in 2005, with essentially no
improvement (and even a slight worsening) over the next decade, reaching 33.4 percent in 2015
(Herrin, 2016).
In 2018, childhood stunting prevalence improved to 29.5 percent, but remains very serious
(Figure 4). Stunting is certainly correlated with wealth status as can be seen by breaking down
childhood stunting by wealth quintile; what is alarming is that the Poorest quintile has a
stunting prevalence of 44.1 percent, close to the national level back in 1989. Prevalence of
severe childhood stunting is also correlated with household wealth; among the poorest quintile,
nearly 15 percent of children under-5 are subject to severe stunting.
9
Figure 3: Stunting prevalence among children under-5, by degree of stunting and wealth
quintile of household (%)
Sources: DOST-FNRI (2022).
Prevalence of anemia and Vitamin A deficiency is most severe among children six months to
one year old.
Deficiency of iron and Vitamin A among children under-5 appears moderate at 13.4 and 15.5
percent, respectively (Figure 5). Iron deficiency though reaches as high as 43.1 percent among
children 6-11 momths old, while Vitamin A deficiency reaches 21.8 percent; and children 1 –
2 years old also suffer very high deficiency prevalence (Figure 5). By age 5 iron deficiency has
fallen to a low 3.7 percent, whereas Vitamin A deficiency falls to the average rate of 14.9
percent.
Iron and Vitamin A deficiency among children under-5 is also correlated with household wealth
quintile.
Prevalence of iron and Vitamin A deficiency among children under-5 is correlated not only
with the age of the child, but also the household wealth quintile (Figure 6). The highest
prevalence of iron and Vitamin A deficiency is among the Poorest quintile, reaching 16.2
percent for iron deficiency, and 22.4 percent for Vitamin A deficiency. Prevalence declines
over the quintiles, falling to 8.6 percent and 5.5 percent among the Richest quintile of iron and
Vitamin A deficiency, respectively.
10
Figure 4: Prevalence of anemia and Vitamin A deficiency among children, pregnant, and
lactating women, Philippines, 2018-19 (%)
Source: DOST-FNRI (2022b).
Figure 5: Prevalence of iron and Vitamin A deficiency among children under-5, by wealth
quintile, Philippines, 2018-19 (%)
Source: DOST-FNRI (2022b).
3.3 Nutrient intake
Household level
The expanded NNS computes energy and nutrient intake by applying the Household Dietary
Evaluation System to convert from raw as-purchased form to cooked; this is combined with
the Philippine Food Composition Table available online, to compute caloric and nutrient
values. Adequacy rate based on share of households with adequate intake, was assessed using
11
the Philippine Dietary Reference Intakes based on caloric and nutrient intakes per capita
(DOST-FNRI, 2022a), respectively referred to as Recommended Energy Intake (REI) and
Estimated Average Requirement (EAR).
Only a minority of households meet dietary norms for intake of energy, iron, and Vitamin A,
though the rate of adequacy rises with wealth quintile.
While a majority of households meet the EAR for protein, only about a fifth meet EAR for
Vitamin A and REI, while only 6 percent meet the EAR for iron (Table 8). For energy and
nutrient intake, adequacy rate tends to rise with wealth quintile, except Vitamin A. In the case
of protein intake, only a minority of households in the poorest quintile meets the EAR, rising
to two thirds in the richest quintile.
Table 8: Share of households whose per capita energy and nutrient intake meet REI and
EAR, Philippines, 2018-19 (%)
Energy Protein Iron Vitamin A
All 21.8 55.1 5.8 22.6
Poorest 20.4 44.9 4.7 21.5
Poor 21.8 51.6 5.6 21.5
Middle 21.5 56.9 5.5 23.1
Rich 22.2 59.0 5.5 23.0
Richest 24.0 66.8 8.2 24.4
Source: DOST-FNRI (2022a).
One reason for the inability to meet dietary intake recommendations is low affordability of
nutritious food, especially fruits and vegetables.
Mbuya et al (2021) computed the cost of meeting minimum requirements for a healthy diet in
2015. The Cost of a Nutritious Diet (CoRD) incorporates dietary guidelines (food “pyramids”),
involving a quantity to be consumed within each food group, to reach nutritional needs and
avoid diet-related non-communicable diseases for an adult male, while adhering to cultural
norms of commonly eaten foods. They find some misallocation of food expenditures based on
2015 food expenditure figures, per adult male equivalent (AME). Filipino households allocate
44 percent more on starchy staples, and 17 percent more for Meat, fish, and nuts, than is
warranted by dietary guidelines (Figure 7). Likewise excess allocation is found for Fats & oils
(17 percent). On the other hand, there is under consumption of Vegetables, Fruit, and Milk and
meat products. Overall however, the daily CoRD of Php 68.2 far exceeds the 2015 average
food expenditure at Php 47.5 per day. This comparison suggests that, even correcting for
misallocation for healthy diets, Filipino households still find a nutritious diet unaffordable.
Part of the reason for unaffordability of a healthy diet is high cost of nutritious food.
World Bank data for 2011 offers valid price comparisons across developing countries,
including the Philippines. Prices are standardized across food groups by normalizing based on
caloric content. Price per calorie for most vegetables and fruits in the Philippines exceeds that
of the average for Southeast Asia (Figure 8); indeed, its cost exceeds that in lower middle
income countries. On the other hand, animal-sourced food groups is cheaper per calorie in the
Philippines than in Southeast Asia.
12
Figure 6: Daily Cost of Recommended Diet and Expenditure per Adult Male Equivalent,
2015
Source: Mbuya et al (2021).
Figure 7: Cost per 1,000 kcal by food category, Philippines and Southeast Asia, 2011
Source: Mbuya et al (2021).
Cereals accounts for the bulk of energy intake, and close to half of protein and iron intake of
households, while Vitamin A is mostly obtained from animal protein sources.
The shares in total energy and nutrient intake by food group are shown in Table 9. Note that
cereals and cereal products are the largest contributors to energy and protein intake, consistent
with the estimates in Figure 7. Of these, Rice & rice products alone contribute the majority of
energy intake (56 percent) and much of the protein and iron intakes (36 and 32 percent,
respectively). Meanwhile, Vitamin A is mostly obtained from animal protein sources; the sum
of Fish & fish products, Meat & meat products, Poultry, Eggs, and Milk, account for 76 percent
13
of Vitamin A intake; the next most significant source of Vitamin A is Vegetables at 18 percent.
The animal protein sources are also the biggest contributor of protein (47 percent) and 2nd-
biggest contributor to iron intake (31 percent).
Table 9: Contribution to energy and nutrient intake of households, by food group,

Philippines, 2018-19 (%)
Cereals & cereal products 68.0 45.4 45.8 3.4
Rice & rice products 56.6 36.5 32.4 0.1
Starchy roots & tubers 0.6 0.3 1.0 0.1
Sugars & syrups 1.8 0.1 0.3 -
Fats & oils 6.4 0.8 1.0 0.2
Fish & fish products 3.5 19.0 10.8 18.6
Meat & meat products 7.5 13.4 11.6 23.2
Poultry 2.0 7.9 3.5 19.3
Eggs 1.6 4.3 3.9 8.3
Milk 1.6 2.4 1.4 6.1
Vegetables 2.2 3.2 8.7 18.4
Fruits 1.2 0.4 1.8 0.8
Others 3.6 2.8 10.2 1.6
Total 100.0 100.0 100.0 100.0
Source: DOST-FNRI (2022a).
The lower the wealth quintile, the greater the dependence on cereals, and the lower the reliance
on animal sources for key nutrients.
The patterns in consumption and expenditure discussed earlier has unavoidable implications
for energy and nutrient intake. The higher wealth quintiles tend to consume less cereals, and
more animal protein sources. Hence contribution of cereals to energy, protein, and iron intake
declines with wealth, while contribution of fish, meat, and poultry to energy, protein, iron, and
even Vitamin A intake increases with wealth (Table 10).
Individual level
Age group is a key determinant of energy and nutrient intake adequacy.

DOST-FNRI (2022b) also presents measures of nutrient intake at the individual level, using
individual consumption data (Table 10). For protein intake, the younger the age group, the
greater the percent adequacy. Meanwhile for energy intake, calorie adequcy is all below twenty
percent, falling to just 9.1 percent for adolescents. Vitamin A intake adequcy is highest for
Infants and preschoolers at 43 percent, declining to 13 percent for Adolescents. Among these
groups, pregnant and lactating women have the lowest levels of adequacy for energy, protein,
iron, and Vitamin A.
14
Table 10: Share of household intake of energy and selected nutrients from cereals and
fish, meat and poultry, by wealth quintile (%)
Poorest Poor Middle Rich Richest
Energy
Cereals & cereal products 75.0 71.5 67.7 64.0 59.7
Rice & rice products 60.9 59.8 57.4 54.0 49.5
Fish, meat, and poultry 7.7 10.4 13.2 16.1 19.1
Protein
Iron
Vitamin A
Rice & rice products 0.1 - 0.2 0.1 0.1
The share of households meeting recommended nutrient intake levels fell from 2015 to 2018-19.
DOST-FNRI (2022b) also provide information the share of individuals meeting nutrient intake
recommendations in 2015; the differences in share are reported in Table 11. Across age groups
and types of nutrients, the share of households meeting nutritional forms has fallen from 2015
to 2018-19, except for energy and protein intake of Lactating mothers. The largest declines
have been registered for Vitamin A intake of Infants & preschool, Schoolage children, as well
as of Pregnant women; also suffering serious decline has been Protein intake of Schoolage
children and of Pregnant women, and Energy intake of Schoolage children.
Table 11: Share of individuals meeting REI and EAR, by age group, Philippines, 2018-19 (%)
Energy (REI) Protein (EAR) Iron (EAR) Vitamin A (EAR)
Infants & preschool 18.4 70.8 4.7 43.2
Schoolage 13.9 67.5 6.3 24.6
Adolescents 9.1 43.9 7.7 13.0
Adults 18.4 45.8 7.8 15.1
Elderly 15.2 27.2 6.4 13.8
Pregnant women 15.1 17.2 7.9 8.2
Lactating mothers 11.7 19.0 7.7 5.5
Sources: DOST-FNRI (2022b).
15
Table 12: Change in share of individuals meeting REI and EAR by age group, Philippines,
2015 vs 2018-19 (percentage points)
Energy (REI) Protein (EAR) Iron (EAR) Vitamin A (EAR)
Infants & preschool -4.8 -5.8 -1.0 -14.2
Schoolage -6.6 -8.6 -0.8 -9.0
Adolescents -1.5 -3.9 -0.8 -5.7
Adults -3.2 -4.3 -1.1 -2.9
Elderly -2.2 -2.6 -0.7 -0.3
Pregnant women -1.0 -6.6 -1.0 -9.3
Lactating mothers 1.9 5.4 -0.6 -3.8
Sources: DOST-FNRI (2022b).
3.4 Government programs
The NNS records household and individual participation in government programs. The
following focuses only on programs recorded in the NNS, that may plausibly affect food
consumption, namely: Pantawid Pamilyang Pilipino Program (4Ps) and the Growth Promotion
Program (GMP).
The 4Ps is a conditional cash transfer program of government in continous implementation
since its pilot phase in 2008. It was finally institutionalized under Republic Act (RA) No. 11310
last 2019. Eligible households are those under the poverty line, as identified under the National
Household Targeting System. Conditionalities for receiving cash include:
• For pregnant women: availment of pre-natal services, give birth in a health facility
attended by a skilled health professional, and receive post-partum and post-natal care;
• For children 0-5: receive regular preventive health and nutrition services
• For children 3 – 5 years: attend daycare or pre-school clases;
• For children 5 – 18 years: attend primary or secondary education classes 85 percent of the
time.
The GMP meanwhile is a nutrition intervention aimed at collection of data on child weight, to
assess interventions to be suggested to parents during counselling sessions. Nutrition workers
will educate parents on the various interventions so as to improve child growth. GMP uses data
as a vehicle for health promotion (DOST-FNRI, 2022).
Majority of households have participated in GMP but only minority in poorest wealth quintile
participate in the 4Ps.
About 71 percent of individuals aged 0 – 71 months participated in GMP. A higher share of
individuals residing in rural areas participated (73 percent). Participation by quintile group is
similar, with a slight tendency for the middle groups to participate more, compared with the
Poorest and Richest groups. The Expanded NNS 2018 also provides data on household
participation in the 4Ps for the poorest quintile (Table 8). Only 43 percent of households in the
Poorest quintile reported participation in the 4Ps. The share is higher among Rural households
in the Poorest quintile, compared with Urban households.
16
Table 13: Share of households and individuals participating in selected government
programs, Philippines, 2018-19 (%)
4Ps households, lowest GMP individuals (0-71
wealth quintile months)
All - 70.6
By place of residence
Urban 33.5 62.7
Rural 45.8 72.9
By wealth quintile:
Poorest 42.7 68.3
Poor - 73.1
Middle - 71.6
Rich - 70.9
Richest - 68.6
Source: DOST-FNRI (2022).
4. Review of literature and methodology
4.1 Conceptual framework
Nutrient intake response begins with a estimating a model of food demand, which is then
converted into intake of energy, protein, iron, and Vitamin A using the FCT of FNRI. Food
demand and nutrient intake is estimated at two levels: first is the household level, which results
in per capita demand and nutrient intake; the second is the individual level, where individual
consumption is the outcome of a sharing rool over the pool of household food demand, also
converted to nutrient intake using FCT. Household demand estimation is in turn is based on
the standard model of a household denoted h, h = 1, 2, ..., N, consuming food items indexed
by i, i =1, 2, ..., M h , corresponding to quantities qih ≥ 0 for each item and in each period. These
quantities must be purchased at market price pi . observing the constraint mh = ∑ i =1 pi qih .
M
Consumption is motivated by maximization of a utility function U h (q1h , q2 h ,..., qMh ) . Under

standard regularity conditions, maximization implies, by the implicit function theorem:
qih = f h ( p1 , p2 ,..., pM , M h ) . (1)
Denote the expenditure share of item i in the budget of household h as sih :
pi qih
sih =
Mh .
An estimate of point-response to a change in consumer price is the elasticity of consumption

with respect to price ε ik where k is an alternative index of i:
∂qih pk
ε ikh − . (2)
∂pk qih
17
The nourishment provided by food is assumed to be measured by a number of indicators j = 1,
2, .., O, with household nutrient intake given by R jh . Supposing nourishment content per unit
of each food item is given by a constant aij , given say by the FCT, then:
R jh = ∑ i =1 aij qijh
N
(3)
We may also define share of item i in overall nourishment R jh as:
ai qih
SRihj =
R jh
.
The demand function in (1) also renders R jh a function of prices and the budget, with its own
elasticity terms, with obvious notation:
ε ikhj = ∑ k sihj ε ik
M
(4)
If the focus is limited to food demand, then under separability conditions, the i items may be
limited to food items, M h total expenditure on food, with M h derived as function of overall
income yh . The specific functional form of f, as well as applications of demand estimation, is
subject to a considerable literature, reviewed briefly below with focus on the Philippines.
4.2 Studies focused on nutrient intake
Quisumbing (1986) provided a comprehensive review of food demand studies up to that year,
showing that demand estimation has been an active field of research among economists. Earlier
studies made extensive use of time series data, e.g. Pante (1971) estimated static and dynamic
demand functions for food, beverages and tobacco, durables, and miscellaneous, using data
from 1949 – 1974. From the mid-1970s onward, demand estimation turned towards cross
section data, available from the Family Income and Expenditure Survey (FIES) under the
National Census and Statistics Office (now absorbed in the PSA), e.g. Goldman and Ranade
(1976). Arboleda (1982) applied the extended Linear Expenditure System (LES) to the 1975
FIES, although he obtained unrealistic results; Canlas (1983) also adopted LES-based approach
and obtained more consistent estimates. Another source of cross-sectional data were the Food
Consumption Surveys (FCS) of the Ministry of Agriculture. Most studies used single-equation
double-log (i.e. Cobb-Douglas-type) approaches, e.g. Bouis (1982), and Regalado (1984).
The NNS data was used for food demand estimation by Quisumbing et al (1988). The study
utilized the FCS components of the NNS for 1978 and 1982. Single equation estimation is
applied following Pitt (1983), with quantity consumped per adult equivalent per commodity
group is the dependent variable. Independent variables are logarithms of prices, together with
total food expenditure, and dummy variables for occupation and year. Owing to numerous zero
consumption observations, parameter estimates are obtained using a Tobit regression.
For a complete simulation, the total food expenditure in turn is estimated using a different data
set, namely the Family Income and Expenditure Survey now conducted by PSA. The overall
demand estimate uses the translog form. Combined with supply side estimates, Quisumbing et
al are able to simulate equilibrium adjustments in energy intake with given policies (e.g.
income transfers and food price subsidies.
18
A similar approach from price and income change to nutrient intake was taken in Orbeta and
Alba (1998), part of a group of studies under the Microeconomic Impacts of Macroeconomic
Adjustment Policies (MIMAP) Project of the Philippine Institute for Development Studies
(Herrin et al , 1992; Orbeta, 1994). Orbeta and Alba (1998) uses the FIES and estimates a full
demand system with eight food items (and one nonfood item). The estimating model is the
AIDS demand system (Deaton and Muelbauer, 1980), in its linear approximate version.
Additional variables were included in the estimation, to control for household demographics
(age and education of the household head, age composition of household members) and region
of residence. To avoid the problem of unconsumed items, the study simply dropped zero
observations. Using the price and income impacts of the 1988-92 tariff reform from a CGE
model, they simulate changes in consumption and ultimately calorie and protein intake. They
find that no only does the tariff reform have a progressive impact on incomes (lower income
groups receive a larger relative income boost), it is even more progressive in terms of
expanding macronutrient availability. Lower income households had a larger increase in
calorie and protein intakes.
4.3 Consumption studies
Also applying the AIDS is Huang and David (1993), this time using United States Department
of Agriculture balance sheet data across countries, one of which is Philippines. Meanwhile the
quadratic extension of the AIDS (Blundell et al, 1993) was applied by Balisacan (1994), also
using the FIES.
An alternative approach outlined by Bouis (1990) and in subsequent papers sidesteps the need
to estimate food demand from consumption data, relying on a demand system based food
characteristics, and calibrated from a few key parameters. He found that during the expansion
phase of the Philippine economy in the 1970s (prior to the crisis of the early 80s), nutrititional
status of low income urban groups did not improve despite lower cereal prices in real terms,
owing to the decline in real wages. In urban areas nutrition likely improved marginally, as as
rural wages declined to a lesser extent.
Several studies tended to focus on specific sub-sectors or locations of food demand, but often
estimating the entire demand system as well. An example of this is Garcia et al (2005), which
focused on fish demand, with data using the FIES. Estimation is conducted in three stages, i.e.
the first stage estimates food expenditure, the second stage estimates fish expenditure, and the
final stage estimates demand for 11 fish types. The final stage adopts the quadratic extension
of the AIDS, with the Heckman correction for zero observations as suggested by Heien and
Wessels (1990).
Another food group receiving attention in demand estimation is vegetables (Mutuc et al, 2007).
Data is obtained from the FIES. They also do a multi-stage estimation, with the first stage
aimed at estimating unit values of demand items, the second stage to model the relationship
between vegetable expenditures and total expenditures, and the third stage to model the
vegetable food demand system using the quadratic AIDS, classified in eleven types. Also
relying on FIES data and using quadratic AIDS is a study on meat consumption (Malabayabas
et al, 2009),
Meanwhile, Lantican et al (2016) used the linear approximate form of AIDS in their study on
rice; BAS (2013) applying the single-equation, double-logarithmic specification, to estimate
demand parameters for selected food commodities. Both these studies drew their data from the
Survey of Food Demand of PSA.
19
Fujii (2016) and Bairagi et al (2022) focus on the location of residence, the former on urban
food demand, the latter on the food basket composition of urban and rural households. Both
make use of the FIES and the quadratic AIDS.
4.4 Role of government programs
As malnutrition indicators for Philippines worsened in the 2010s, PIDS revived emphasis on
the topic by a series of studies related to nutrition programs and policies. The series kicked off
in 2016 with Herrin (2016), which focused on stunting among preschool children. The paper
underscored importance of maternal nutrition and health, and key interventions, which include
infant and young child feeding and micronutrient supplementation, both of which relate to
nutrient intake.
Ulep et al (2021) report a statistical study based on a data set with high stunting prevalence
(38.5 percent), with prevalence rising to 45 percent among the poor (versus 32 percent for non-
poor). Over half of stunting prevalence is due to maternal factors, while dietary diversity and
iron supplementation account for 12 percent and 5 percent of child stunting, respectively.
Another set of studies focus on nutrition program implementation. A follow-up study of Herrin
et al (2018) focused on LGUs as the frontline in service delivery to end child stuning in the
Philippines. It affirms the effectiveness of “The First 1000 days” strategy, which include
complementary feeding and micronutrient supplementation. LGU implementation should be
accompanied by monitoring and evaluation towards consistent service delivery.
Dacuycuy et al (2019) as well as Abrigo and Tam (2019) as well as offer case studies of LGU
nutrition programs. The latter recommends based on the experience of Zamboanga del Norte
that LGUs need to focus limited resources on high-impact interventions, including infant and
young child feeding, and micronutrient supplementation among at-risk children.
5. Empirical strategy for household demand
5.1 Data source
The method of this study combines the various techniques that have been implemented in past
studies. The data source for household food demand is the FNRI, using the FCS for 2013, 2015,
and 2018, using information on household food intake, food cost, demographic characteristics,
year, and participation in selected government programs. The FCS also includes data on
occupations, coded as follows:
0 - Special Occupations
1 - Official of Gov't, Corporate Exec, Mngr, Managing Prop, Supv.
2 - Professional
3 - Technicians & Associate Professionals
4 - Clerks
5 - Service Workers & Shop and Market Sales Workers
6 - Farmers, Forestry Workers & Fishermen
7 - Craft & Related Trades Workers
8 - Plant & Machine Operators and Assemblers
20
9 - Elem. Occupation: Laborers & Unskilled Workers
10 - Housekeeper, Pensioner, Student & No Occupation
11 - No PSOC
99 - Not applicable (<10y)
9999 - Missing
As argued by Bouis, surveys of food intake based on direct measurement is a better source of
data than food expenditure surveys, which are prone to measurement errors that end up
exaggerating expenditure elasticities of demand. For instance in the Philippines, the calorie-to-
expenditure elasticity falls within the 0.40 – 0.52 range when using food expenditure surveys,
but declines to 0.13 to 0.17 when using food intake surveys (Bouis et al, 1992).
Based on the categories of Table 7, the estimation is based on ten food groups, as follows:
1. Rice - Rice & rice products
2. Other cereals – Cereals & cereal products aside from Rice & rice products
3. Fish – Fish & fish products
4. Meat – Meat & meat products
5. Poultry – Poultry and Eggs
6. Fruit
7. Vegetables
8. Dairy – Milk
9. Fats & oils
10. Other food – Starchy roots & tubers; Sugar & syrups; Others
The unit values of the foregoing food groups (obtained by dividing food group expenditure by
food intake) is taken as the food group price. Expenditures shares for the full data set (2013,
2015, 2018) at the sample mean is summarized in Table 14; also shown are unit values, and
the share of households who have positive consumption over the food group. Rice (and rice
products) accounts for a quarter of food expenditure; the next largest food group by expenditure
share is Fish at 18 percent; the food group with lowest share is Fruit at 1.5 percent. On a per
kg basis the most expensive food group is Other cereals, followed by Fats & oils, and then
Meat. The last column shows that nearly all households consume Rice, while nearly nine-tenths
consume Vegetables. However, zero consumption accounts for a substantial share of the the
rest of the food groups (except Other food). The least consumed item is Fruit at just 21 percent
of households; only 35 percent consume Dairy, and less than half consume Meat.
21
Table 14: Mean values of selected variables, by food group, FCS
Expenditure shares Unit values (Php per Share of consuming
(%) kg) households (%)
Rice 25.4 47.4 96.8
Other cereals 6.6 322.8 59.7
Fish 18.0 115.7 78.4
Meat 11.2 167.0 49.1
Poultry 8.5 138.0 55.7
Fruit 1.5 58.8 21.0
Vegetables 9.8 89.8 88.9
Dairy 3.6 88.8 35.1
Fats & oils 2.3 140.4 82.9
Other food 13.3 133.1 99.2
5.2 Case of purchased food groups
The regression model is the quadratic AIDS, specified for the food demand system. Ignorinthe
problem of non-purchases, the system is specified as follows:
N
γ 0i + ∑ γ ij ln p jh + βi ln( M h Ph ) + λi [ln( M h P h )]2 + ∑ δ ik zkh + ε ih
sih =
j k
(1)
Here the i-term indexes food groups, while M h denotes food expenditure. Ph is the Stone price
index ( Ph = ∑ s jh ln p jh ); and zkh denotes other control variables corresponding to household

j
demographics, location, and survey period. Parameters to be estimated are α i , γ ij , βi , λi , δ i ,

while ε ij ~ N (0,1) is the error term. The expected value of (1) drops the error term and leaves
a right hand side expression denoted here as QuAIDS.
Utility maximizing subject to an expenditure constraint entails the following:
γ ij = γ ij
(symmetry);
∑ γ 0i 1;=
= ∑ γ ij 0;=
∑ βi 0;=
∑ λi 0
i i i i (adding up).
If so then the parameters of one of the Nth equations can be recovered by estimating the N-1
equations. The symmetry restriction is handled here using the seemingly unrelated regression
technique (SUR), using the appropriate set of commands in STATA.
22
Elasticity terms are evaluated at QuAIDS. Expenditure elasticity for food group i,denoted η mih ,
is computed by using ln sih = ln pi + ln qih − ln mh , and performing logarithmic differentiation on
Equation (1):
1   ln M h  
η mih =
1 +  βih + 2λih  
sih   Ph   (2)
A similar technique implies the own- and cross-price elasticity terms, denoted ε ij . In the
following, uij is the Kronecker unit, with u=
ii 1, u=
ij 0, i ≠ j :
1   M  
η pijh =−uij + γ ijh − s jh  βi + 2λi ln   
sih    P   (3)
Elasticities should also adhere to the following under utility maximization:
∑ s ηmih ih =1
i ;
∑η p ijh + η mih =
0
j
.
This pertains to household demand as a function of prices and food expenditure. Household
demand needs to be related to household income in order to make linkages to macro variables
such as GDP. For this purpose, the FIES is utilized. Denote consumer price index for food and
for all items as CPIFh , CPI h , respectively. Least squres regression is applied to estimate a
double-log equation with the usual error term:
mh y
ln α 0 + α1 ln h + ε mh
=
CPIFh CPI h
Taking expectations yields
ln m=
h a0 + α1 ln( yh ) , a0 =
α 0 + ln CPIFh − α1 ln CPI h . (4)
Case of zero purchases for some food groups
Some food groups of policy interest may not be purchased at all by some households, i.e. not
all households consume rice (they be maize and wheat eaters). To account for zero purchases,
the Heckman selection model posits a latent random variable yij such that qih is observed
conditional on wih > 0 . In turn, wih is determined by independent variables denoted x1 , x2 ,... ,
and an error term ν ih such that:
µ0i + ∑ µil xlh +ν ih , ν ih ~ N (0,1)

wth =
l . (5)
Crucially, the error terms are correlated with a correlation term ρi > 0 , implying an omitted
variable problem in directly estimating (1).
23
Denote the standard normal density, and cumulative normal density functions, by φ , Φ ,
respectively. Taking the expected value of (5), and using the formula for moments of a
truncated bivariate normal distribution (Greene, 2017), letting wih = E [ wih ] :
  
E [ qih |=
wih > 0] E  qih | ν ih > −  µ0i + ∑ µil xlh  
  l 
φ ( wih )
= QuAIDS + ρi
Φ ( wih )
. (6)
The ratio term is shortened to IMR, the Inverse-Mills ratio. Estimation follows the procedure
outlined by Heckman’s two step estimator: first is to estimate (5), using probit regression,
imposing wih to be a binary variable with wih > 0 when qih is observed, 0 otherwie. The
regression is then used to obtain predicted values wˆ ih = ∑ µˆ ilh xilh , where “^” denotes the value
l
of the estimator. Then obtain IMR (rˆ) , to augment the original QuAIDS as in (6), to be
estimated in the usual manner.
Incorporating food price and expenditure in the selection model
The list of independent variables has two options: one is to include only demographic variables;
another is to include demographic variables, together with prices and expenditure. Recent
examples of the former are Mariscal and Werner (2018), and Lokuge et al (2019). For the latter
there Koeshandrajana et al (2021), which includes expenditure, and Ashidigbi et al (2019),
which includes prices and expenditure. This study adopts the latter option, for which latent
variable model also includes price and expenditure varaibles. For instance, the marginal meat-
eater (i.e. purchasing meat only once a month) may cease meat consumption entirely if meat
price is sufficiently high, or if income drops sufficiently low. The validity of including price
changes can be checked by examining the coefficients of these price variables in the first stage
model.
In this case the marginal effect of a change in price and income on consumption consists of a
direct effect captured by QuAIDS, and an indirect effect through the latent variable wih . The
extant literature which adopts this first stage option appears not to have acknowledged the
presence of an indirect effect.
It should be noted that, conditional on strictly positive consumption, the QuAIDS is
consistent with utility maximization subject to constraints; hence, with this proviso, the usual
restrictions will apply from the combination of differentiability, symmetry, and zero degree
homogeneity in prices and income. However, considering the prior stochastic selection process
into strictly positive consumption, these restrictions may be inapplicable once the inverse-Mills
correction is applied.
Suppose (5) is specified as follows (non-price and expenditure terms suppressed):

µoi + ∑ µij ln p jh +µ1i ln mh +ν ih
wih =
j
(7)
24
Performing the differentiation:
∂IMR ( wih )
− µ1i IMR ( wih )  wih + IMR ( wih )  ;
=
∂ ln mh
∂IMR ( wih )
− µij IMR ( wih )  wih + IMR ( wih )  .
=
∂ ln p jh
The adjusted elasticity formulas are therefore:
1   ln M h    − ρi 
η mih =
1 +  βih + 2λih   +   µ1i IMR ( wih )  wih + IMR ( wih ) 
sih   Ph    sih  (8)
1   M  
η pijh =−uij + γ ijh − s jh  βi + 2λi ln    +
sih    P  
 − ρi 
  µij IMR ( wih )  wih + IMR ( wih ) 
 sh  (9)
Explanatory variables for the first stage probit estimation are as follows: prices (in logarithms);
total food expenditure (in logarithms); and dummy variables for region and survey period
(omitting region BARMM and 2013).
For the second stage, estimation is based on (6), with z-variables consisting of the following:
• Household demographics –educational attainment of mother, age of mother, dummy
variable for skilled occupation of household head
• Household asset index – FCS wealth quintile (numbered 1 to 5, in logarithms)
• Region dummies (BARMM omitted) and survey year dummies (2013 omitted)
For households without mothers, household head age and education are used in the estimation.
Skilled occupations cover codes 1 to 5, 7, and 8.
Simulations of nutrient intake and individual demand
Household food demand is readily translated to energy and nutrient intake using nutrient
contents derived from FCT. The following applies to energy intake Eh , with corresponding
energy content aih per unit of food group i, but can readily be applied to any selected nutrient
intake:
N
Eh = ∑ aih qih
i =1
Denote “%” as the percentage operator, e.g. dq q = %q ; this implies:
a q 
% Eh = ∑  ih i %qi 
i  Eh .
The fractional term is simply the share of food group i in energy intake of household h. In turn,
household demand change for a given set of price and income changes is approximated using
25
the elasticity estimates; note though the smaller the price and income change, the better the
approximation, as the elasticity formulas are based on point derivatives.
%qi | dp j = % p jη pij
%qi | dmh = %mh (η mih )
In addition to changes in food price and in nominal income, adjustment should also be made
for changes in overall CPI, in accordance with Equation (4) under logarithmic differentiation:
%mh =α1 % yh + [ %CPIF − α1 %CPI ]
Lastly, for individual demand: One option is direct estimation of individual food demand as in
equation (1) as suggested in Deaton (2018). Another option, taken in this paper, is to perform
a double-log estimation using least squares, with consumption of group i for a given household
member type as the dependent variable, and that the household.
6. Results of demand estimation
6.1 The selection model for household food demand
Price and expenditure variables are mostly statistically significant in predicting positive
consumption.
The independent variables of the selection model are the prices (in logarithm), food expenditure
(in logarithm), region dummies, and survey year dummy. Coefficients of price and expenditure
variables from the probit regression are shown in Tables 15 and 16. The best-fitted model based
on the latter is Rice, and least well fitted is Fish. The rest of the regressions have goodness-of-
fit between 0.08 to 0.63 based on psuedo-R2.
Table 15: Coefficients of price and expenditure variables, probit regressions

Explanatory Dependent variable (Binary)
variables Rice Other food Vegetables Fats & oils Fish
Rice -0.334 -0.425 -0.070 0.117 -0.158
Other cereals 1.103 -0.051* -0.056 0.039 -0.035
Fish 0.004* -0.060 *
-0.019 *
0.028* -0.536
Meat -0.109* 0.016 0.010 *
-0.222 0.027*
Poultry -0.130 -0.069* -0.064 -0.221 -0.008*
Fruit -0.127 -0.028 *
-0.020 *
0.033 -0.004
Vegetables 0.159 -0.119 -0.047 -0.057 0.086
Dairy -0.042* -0.064* -0.043 -0.028 -0.008*
Fats & oils -0.094 -0.172 0.015* -0.479 -0.042
Other food 0.109 -0.205 -0.114 0.007* -0.035
Expenditure 0.603 0.697 0.633 0.636 0.603
Pseudo-R2 0.634 0.191 0.133 0.152 0.634
*
Not significant at 5 percent level.
Source: Author’s calculation.
Price and expenditure variables are all statistically statistically significant in predicting positive
consumption of Poultry and Meat food groups; only one price variable is not statistically
significant in the probit regression for Other cereals and Fruit (Poultry for both). The majority
26
of price and expenditure coefficients are statistically in all equations. Own-price coefficients
are negative and statistically significant, as are expenditure coefficients.
Table 16: Coefficients of price and expenditure variables, probit regressions

Explanatory Dependent varaible
variables Other cereals Poultry Meat Fruit Dairy
Rice 0.122 0.203 0.264 0.140 -0.035*
Other cereals -0.583 0.088 0.054 -0.006 0.028
Fish 0.015 0.041 0.109 0.090 0.066
Meat -0.053 -0.156 -0.724 0.033 -0.065
Poultry 0.007* -0.767 -0.082 -0.002* -0.049
Fruit 0.008 0.058 0.065 -0.792 0.519
Vegetables 0.057 0.101 0.180 0.033 0.519
Dairy -0.099 -0.054 -0.050 -0.037 -0.830
Fats & oils 0.086 -0.033 0.058 0.049 0.002*
Other food 0.097 0.078 0.060 -0.022 -0.034
Food expenditure 0.421 0.505 0.736 0.456 0.663
Pseudo-R2 0.143 0.106 0.200 0.134 0.166
*
Not significant at 5 percent level.
6.2 Food demand estimation
Parameters of the household food demand system
For the estimated demand model, price and expenditure terms are mostly statistically
significant; government programs have a mixed effect on expenditure shares.
The results of the demand system seemingly unrelated regression are relegated to the
Appendix. STATA reports an analogue for adjusted R2; goodness-of-fit using this measure is
highest for Rice (0.37) and lowest for Fruit (0.03); goodness-of-fit is mostly below 0.20, which
is low but not unsual given the cross-section nature of the data.
Expenditure terms are positive, as well as the squared expenditure terms. Moreover,
expenditure terms are all significant. Hence, Engel curves slope upward and become steeper.
Price terms are also significant, with some exceptions: dairy with the most number of non-
significant price coefficients. Participation in 4Ps and GMP has positive effect on expenditure
shares of some food groups, but negative on others; statistically insigificant coefficients for
Fruits and Vegetables. Finally, household wealth reduces expenditure share of rice, fish,
vegetables, but raises that of the other food groups.
Income elasticities of household food demand
Most food groups are normal goods, among which income elasticities of food demand
range from 0.73 (Poultry) to 1.75 (Other cereals).
The food expenditure elasticities at the sample mean, both with selection effects, and the
original QuAIDS formula, are shown in Table 17. Also shown are the income elasticities, using
the estimated elasticity of food expenditure to income (using the double-log specification) of
0.5714. Poultry and Rice have nearly identical income elasticities (0.73 and 0.75, respectively).
27
Table 17: Elasticity of household consumption with respect to food expenditure and
income, with and without selection effects
With selection Without selection
Food expenditure Income elasticity Food expenditure Income
elasticity elasticity elasticity
Rice 1.32 0.75 1.19 0.68
Other cereals 3.06 1.75 3.12 1.78
Fish 1.71 0.98 1.49 0.85
Meat 2.13 1.22 1.98 1.13
Poultry 1.29 0.73 1.42 0.81
Fruit 1.72 0.98 2.98 1.70
Vegetables 1.91 1.09 1.48 0.85
Dairy 2.88 1.65 3.09 1.77
Fats & oils 2.96 1.69 3.09 1.77
Others -4.36 -2.49 -3.56 -2.04
Source: Author’s calculation
Fish and Fruit demand are both close to unit elastic, while Vegetables income elasticity is 1.09.
Note that for the majority of normal goods, incorporating selection effects reduces high values
of income elasticity from QuAIDS (i.e. Other cereals, Poultry, Dairy, Fats & oils, and
especially Fruit). The weighted average of income elasticities with and without selection effects
is very similar (0.56 and 0.55, respectively).
Price elasticities of household food demand
For a majority of the food groups, demand is inelastic to own price; the most purchased food
groups tend to be complementary goods except Other food.
Price elasticity estimates are shown in Table 18, using the full elasticity formula with selection
effects. Own-price elasticities are all negative, and are inelastic (absolute value below unity)
except for Other cereals, Fruit, and Other food. Rice, Fish, Meat, and Poultry tend to be
complements with each other and the other food groups, though some obvious substitution
effects are confirmed, such as Rice and Other cereals, and Other food with the rest of the food
groups.
Table 18: Price elasticities of household demand, with selection effects

Food group Price variable
Rice Other cereals Fish Meat Poultry
Rice -0.694 0.290 -0.180 -0.140 -0.124
Other cereals 0.217 -1.814 -0.260 -0.130 -0.031
Fish -0.355 -0.152 -0.829 -0.087 -0.110
Meat -0.211 -0.098 -0.108 -0.817 -0.258
Poultry -0.232 0.028 -0.080 -0.171 -0.614
Fruit -0.004 -0.007 0.155 0.020 -0.034
Vegetables -0.310 -0.121 -0.034 0.059 -0.011
Dairy -0.059 -0.084 -0.092 -0.159 -0.190
Fats & oils -0.319 -0.042 -0.140 -0.098 -0.155
Other food 0.762 0.888 0.227 0.441 0.725
28
Fruit Vegetables Dairy Fats & oils Other food
Rice -0.049 -0.101 -0.056 -0.090 0.071
Other cereals -0.139 -0.200 -0.130 -0.145 -0.440
Fish -0.057 -0.029 -0.095 -0.123 -0.038
Meat -0.078 0.018 -0.152 -0.119 -0.174
Poultry -0.051 0.038 -0.074 -0.069 -0.076
Fruit -2.267 0.038 -0.082 -0.034 0.653
Vegetables -0.039 -0.959 -0.062 -0.071 -0.209
Dairy -0.303 -0.179 -0.894 -0.100 -1.070
Fats & oils 0.029 -0.175 -0.087 -0.377 -1.562
Other food 0.907 0.683 0.786 0.829 -1.344
Table 19 shows the alternative elasticity calculation based only on the second stage QuAIDS,
ignoring the secondary effects of price and expenditure on selection. Compared with the
adjustments for expenditure elasticity, selection effects appear to have a smaller influence on
price elasticity, although the Table 18 values tend to check some large own-price elasticities in
Table 19, e.g. Other food, Dairy, and most especially, Fruit.
Table 19: Price elasticities of household demand, without selection effects

Food group Price variable
Rice Other cereals Fish Meat Poultry
Rice -0.622 0.052 -0.181 -0.116 -0.096
Other cereals 0.233 -1.890 -0.258 -0.137 -0.031
Fish -0.271 -0.133 -0.541 -0.102 -0.106
Meat -0.264 -0.109 -0.129 -0.674 -0.242
Poultry -0.180 0.051 -0.070 -0.211 -0.811
Fruit 0.382 -0.024 0.401 0.111 -0.039
Vegetables -0.264 -0.084 -0.021 0.052 0.032
Dairy -0.070 -0.075 -0.071 -0.179 -0.206
Fats & oils -0.293 -0.034 -0.134 -0.146 -0.202
Other food 0.277 0.829 0.159 0.459 0.646
Fruit Vegetables Dairy Fats & oils Other food

Rice -0.022 -0.135 -0.047 -0.070 0.047
Other cereals -0.138 -0.192 -0.143 -0.134 -0.427
Fish -0.054 -0.075 -0.091 -0.101 -0.019
Meat -0.091 -0.017 -0.142 -0.130 -0.186
Poultry -0.036 0.064 -0.088 -0.078 -0.056
Fruit -4.445 0.128 -0.184 0.101 0.594
Vegetables -0.026 -0.927 -0.033 -0.081 -0.133
Dairy -0.140 -0.016 -1.154 -0.099 -1.081
Fats & oils 0.036 -0.187 -0.093 -0.480 -1.561
Other food 0.875 0.548 0.713 0.633 -1.577
29
Individual consumption elasticities
The estimated elasticities of individual to household consumption are summarized in Figure 8.

Elasticities are all positive, but far below unity, implying a large departure from a constant
sharing rule. For children under-5 the elasticities are highest for Dairy, and Other cereals; for
pregnant and lactating women the elasticities are highest for Other cereals, followed by Fish.
Figure 8: Individual-to-household consumption elasticities, by household member type
6.3 Elasticities of energy and nutrient intake
Income elasticities
Income elasticities for energy and nutrient intake are very high, in the range of 0.80 to 1.0.
The resulting total elasticities of intake of calories, protein, iron, and Vitamin A, to changes in
income are summarized in Figure 9. Note that incorporating selection effects has almost no
quantitative influence on the energy and nutrient intake elasticities. The estimates for energy
intake are far above the recommended range of 0.08 to 0.14 range suggested by Bouis and
Haddad (1992). The possible reasons for breaching this range are taken up in the discussion of
study limitations.
30
Price elasticities
Elasticity of energy and nutrient intake with respect to prices of the major food groups is shown
in:
Figure 9: Intake elasticities with respect to income, energy and selected nutrients
The food group with the largest price elasticity for energy and protein intake (in absolute terms)
is Rice; a one-percent increase in the price of rice is enough to reduce energy intake by 0.39
percent, protein intake by 0.31 percent, and iron intake by 0.22 percent. Even Vitamin A intake
declines by 0.22 percent, even though rice is not a source of Vitamin A, mainly due to
complementarities with other food groups that contribute most to Vitamin A intake. The food
group with the largest price elasticity for Protein intake (aside from Rice) is Fish, followed by
Other food. For iron the largest price elasticities are for Other food, followed by Fish. Lastly
for Vitamin A the largest price elasticities are for Poultry, Meat, and Fish.
Table 20: Intake elasticities with respect to price, energy and selected nutrients
With respect to price of: Energy Protein Iron Vitamin A
Rice -0.39 -0.31 -0.22 -0.22
Other cereals -0.01 -0.04 -0.07 -0.12
Fish -0.17 -0.25 -0.17 -0.22
Meat -0.15 -0.17 -0.13 -0.25
Poultry -0.09 -0.12 -0.05 -0.26
Fruit -0.03 0.00 -0.01 -0.08
Vegetables -0.07 -0.04 -0.06 -0.17
Dairy -0.04 -0.04 0.01 -0.13
Fats & oils -0.06 -0.04 0.00 -0.08
Other food -0.22 -0.18 -0.25 -0.20
31
7. Simulated impacts of recent economic shocks: household level
7.1 Set-up of scenarios
Covid19 with and without Bayanihan social protection
Quantifying the impact of Bayanihan social protection programs involves an adjustment of the
income shock term in the previous simulation, which adopts the GDP contraction of 2020
without accounting for transfers schemes to ameliorate loss of purchasing power of households.
This corresponds to a nominal GDP contraction of 7.9 percent, an overall inflation rate of 2.4
percent, and a food inflation rate of 2.9 percent.
The household income shock with Bayanihan programs is estimated to be 5.63 percent higher
than the baseline of without Bayanihan, based on the figures shown in Table xx, computed
from Annual Poverty Indicators Survey (APIS) 2020. The average is compared with the
estimated average income of SAP recipients reported in DWSD-WFP (2022). The COVID19
scenarios are applied to the elasticities evaluated at the sample mean of the FCS.
Table 21: Coverage and benefit estimates for Bayanihan social protection programs, 2020
Coverage of Amount per
households (%) household (Php)
SAP 41.9 6,176
DOLE TUPAD 7.0 3,987
DOLE CAMP 6.0 5,383
DOLE AKAP 1.0 10,441
DA Rice Farmers Assistance 2.9 4,967
DSWD Relief Assistance 22.0 940
Other relief assistance from GPH 85.5 1,712
TOTAL (weighted average) na 5,109
Sources: Cho and Johnson (2022); DSWD and WFP (2022).
Inflation with growth in 2022
Between October 2021 and October 2022, the overall inflation rate was 7.67 percent; this was
exceeded by one of its biggest components, namely Food price inflation, at 9.79 percent.
Alongside this was nominal GDP growth from 2nd quarter 2021 to 2nd quarter 2022 equal to
11.26 percent. Likewise these price and income shocks are applied to the elasticities evaluated
at the sample mean of the FCS.
7.2 Results
COVID19 scenario
The economic shock of COVID19 caused energy and nutrient intakes to decline, with the
severest contraction in Vitamin A intake.
Projected changes in energy, protein, iron, and Vitamin A intake within the COVID19 period
(combining moderate inflation with severe income contraction) are shown in Table 22. Energy
and nutrien intakes fall; the declines in energy, protein, and iron intake lie within the 6-7 percent
range; however that of Vitamin A is 12 percent. The main major sources of decline in energy,
32
protein, and iron, are Rice, Other cereals, Fish, and Meat, while that of Vitamin A decline are
consumption of Vegetables, Fish, and Meat. At the individual level, declines for children under
six are about half those of the household level, while that of pregnant and lactating mothers
lies within the 1 – 2 percent range.
Table 22: Changes in intake of energy and selected nutrients, COVID19 2020 scenario,
household and selected members (%)
Total household intake -6.98 -6.70 -6.14 -12.11
By food group
Rice -2.42 -1.55 -1.41 0.00
Other cereals -2.68 -2.15 -3.06 -0.80
Fish -0.41 -2.14 -1.25 -2.18
Meat -0.93 -1.58 -1.46 -2.87
Poultry -0.25 -0.76 -0.49 -1.94
Fruit -0.34 0.00 -0.69 -0.25
Vegetables -0.37 -0.60 -1.36 -3.04
Dairy -0.40 -0.45 -0.29 -1.50
Fats & oils -1.25 0.00 -0.23 -0.04
Others 2.06 2.52 4.08 0.51
By individuals:
Children under-5 -3.11 -3.73 -3.06 -5.63
Pregnant and lactating mothers -1.16 -1.58 -1.22 -1.41
Source: Authors’ calculation.
COVID19 with Bayanihan
The Bayanihan social protection schemes ameliorated the negative energy and nutrient intake
impact of COVID-19 by one-half or more.
The impact of the COVID19 with Bayanihan is shown in Table 23. As the average income
transfer falls short of the average economic contraction due to COVID19, the decline in energy
and nutrient intake cannot entirely be reversed. Nonetheless the decline is much less than in
the previous scenario, at around 2.5 – 2.6 percent for energy, protein, and iron, and 6.5 percent
for Vitamin A. Although the fall in income is 71 percent smaller than in the previous scenario,
the decline in energy and nutrient intake is smaller by only 47-65 percent, showing that even
partial reversal of nutrient and energy intake declines is a highly expensive proposition for
government.
33
Table 23: Changes in intake of energy and selected nutrients, COVID19 with Bayanihan,
household and selected members (%)
By food group
Rice -0.01 -0.01 -0.01 0.00
Other cereals -1.56 -1.25 -1.78 -0.46
Fish -0.22 -1.14 -0.66 -1.16
Meat -0.42 -0.70 -0.65 -1.28
Poultry -0.10 -0.31 -0.20 -0.78
Fruit -0.28 0.00 -0.56 -0.20
Vegetables -0.23 -0.38 -0.86 -1.91
Dairy -0.25 -0.28 -0.18 -0.94
Fats & oils -0.63 0.00 -0.11 -0.02
Others 1.23 1.50 2.43 0.30
By individuals:
Children under-5 -1.83 -2.24 -1.85 -3.49
Inflation with growth in 2022
Despite rapid economic growth in 2022, the accelerated inflation in food prices in the period
caused a slight to moderate decline in energy, protein, and iron intake, and a moderate decline
in Vitamin A intake.
The rapid economic recovery from 2021 onward has raised hopes of reversing the projected
decline in energy and nutrient intake from the COVID19 pandemic. Unfortunately, the
accelerating food price inflation that happened to coincide with recovery is enough to sustain
the decrease energy and nutrient intakes two years after the pandemic. Fortunately, the declines
are quite small, in the range of 0.6 to 1.3 percent for energy, protein, and iron. The decline in
Vitamin A is however more serious, at 5.4 percent; the fall is mostly due to price increases for
Meat, Poultry, and Vegetable. Another source of concern is the unusually high reduction in
nutrient intake of Children under-5; this is largely due to the high Dairy price increases.
Table 24: Changes in intake of energy and selected nutrients, Inflation with growth
scenario, household and selected members (%)
By food group
Rice 0.94 0.60 0.55 0.00
Other cereals -1.39 -1.12 -1.59 -0.41
Fish -0.18 -0.96 -0.56 -0.98
Meat -0.43 -0.73 -0.67 -1.33
Poultry -0.13 -0.41 -0.26 -1.04
Fruit 0.07 0.00 0.15 0.05
34
Vegetables -0.18 -0.29 -0.66 -1.47
Dairy -0.15 -0.17 -0.11 -0.56
Fats & oils -0.59 0.00 -0.11 -0.02
Others 1.48 1.81 2.92 0.37
By individuals:
Children under-5 -1.01 -1.33 -0.99 -2.02
8. Conclusion
8.1 Contributions and limitations of the study
Analysis of changes in nutrient intake as a response to economic shocks were common in the
1980s and 1990s, but recently economic shocks from 2020 onward have been relatively
understudied.
This study is an attempt to remedy this gap in the literature by updating demand elasticity
estimates, combinining it nutrient content information, and applying the resulting model to
simulate the impact of the recent COVID19 household income shock, and the 2021-22 inflation
spikes.
Recent food and nutrition surveys conducted by the Philippine government are a valuable
source of information about household behavior and the impact of economic shocks.
The Philippines has regularly conducted FCS based on food intake and direct weighing;
however a review of past studies found that its use in food demand estimation seems to have
ceased in the 1980s and 1990s, largely displaced by income and expenditure survey data. This
paper shows that FCS is a viable source of data for such modeling approaches, in particular the
flexible functional form with variable expenditure elasticities known as the AIDS. Demand
modeling in turn is shown to be an invaluable tool for evaluating the impact of economic
shocks, and therefore of offsetting policy measures.
Selection effects are important determinants in consumption and should be taken into account
in calculating consumption response to prices and income.
In using FCS (and household survey data for that matter), some technical issues need to be
addressed. One emphasized in this paper is sample selection effects, which lead to zero
consumption of nutritionally meaningful food groups for some sample households. When the
selection model includes price and expenditure variables, as is plausible, then the standard
elasticity formulas need to be adjusted. This paper outlines the empirical strategy for making
the appropriate adjustment.
The current study is limited by its selection of commodity aggregates in the estimation.
Bouis and Haddad (1992) mention another source of measurement error in estimating calorie
income elasticities (and by extension, other nutrient elasticities), namely the change in
composition of a food group towards higher quality items (in terms of taste and packaging or
presentation), but with lower energy and nutrient content per unit weight. They find other
sources of measurement error more important, but that is because their specific data set seems
to have stable food group characteristics, e.g price per kg for rice and corn vary little by
household income. In contrast, we find in the FCS data for this study, a correlation of 0.24
between total household expenditure and unit value of rice; a correlation of 0.19 with that of
unit value of other cereal; 0.10 with that of unit value of meat; and 0.09 with that of unit value
35
of poultry. The correlations are all statistically significant at 5 percent level. One way to address
this is to disaggregate beyond the ten food groups adopted here; however the trade off is to
aggravate the selection problem already discussed. Addressing this issue is beyond the scope
of this paper; further research is needed to more thoroughly address issues related to
aggregation.
Notwithstanding these technical issues, the study was nevertheless able to determine that the
that the COVID19 social protection programs played a significant role in preventing further
deterioration in nutrient intakes and worsening of malnutrition. Moreover, the study also found
that, notwithstanding rapid economic growth, the recent inflation episodes pose a major threat
to nutrient intakes and nutrition security.
8.2 Implications for policy
Income policies in the form of targeted cash transfers are an important, albeit expensive way to
counter adverse nutrition impacts of economic contraction.
The simulation analysis conducted in this study shows that the Bayanihan programs were an
effective way to counter the adverse nutrition impacts of economic contraction brought about
by the COVID19 pandemic. However, these programs are very expensive and are not
sustainable. One way to keep the fiscal cost down is to target cash transfers to the most
vulnerable groups, i.e. the poorest households.
Ameliorating the impact of price increases during inflation episodes should assume priority in
policy research and response.
There is a tendency to downplay welfare impacts of price increases during periods of rapid
economic growth, believing the latter more than cancels out the former. This is not always the
case though, as the analysis of the 2021-22 episode done in this paper shows. Hence each
inflationary episode should be examined individually to determine the net effect of economic
changes on household food security. If the net impact is negative and quantitatively significant
then framing the appropriate policy response assumes high priority.
Cost-effective policies to improve food affordability, such as trade liberalization, should be
pursued aggressively when nutrition security is under threat.
A common tendency of policymakers is to favor price subisidies in order to keep food
affordable. Consumer subsidies are however financially unsustainable; hence, for instance, the
tantalizing promise of “Php 20 rice” could not in fact be delivered owing to high fiscal cost,
and the opportunities foregone from an expensive food subsidy scheme. Much more cost-
effective are trade liberalization measures. The food groups that contribute most to energy,
protein, and micronutrient intake of Filipinos, namely Rice, Other cereals, Fish, Meat, and
Poultry, are all produced under high levels of trade protection against cheaper imports. This
has the unfortunate consequence of reducing affordability of nutrient-rich foods. The sooner
the government dismantles high tariffs and overly strict (and often arbitrary) application of
sanitary and phytosanitary standards on these major consumer goods, the more affordable these
items become especially to the poor.
36
Productivity improvement along the value chain for rice, animal protein, and vegetables, are
have the potential for simultaneously raising incomes and making food more affordable to the
poor.
Productivity improvement, beginning from farm production , and extending through out the
food value chain, has the potential of boosting competitiveness against cheap imports, raising
incomes of food producers , processors, and distributors. By lowering production and logistics
cost, productivity improvements also tend to make food more affordable. Identifying the right
investments, and sutiable actors along the value chain, to realize these productivity
improvements, poses a real challenge to agro-industrial policy; however the potential benefits
are too large to ignore this particular approach to food policy.
9. Bibliography
Abrigo, M., and Z. Tam. 2019. Situational analysis of distal factors affecting early childhood
care and development in the first 1000 days of life in Zamboanga del Norte. Discussion
Paper Series No. 2019-28. Quezon City: PIDS.
Arboleda, H. 1982. A Study on the Consumption and Saving Patterns of Filipino Families with
the use of the Extended Linear Expenditure System. Master's thesis. The Hague,
Netherlands: Institute of Social Studies.
Bairagi, S., Y. Zereyesus, S. Baruah, and S. Mohanty. 2022. Structural shifts in food basket
composition of rural and urban Philippines: Implications for the food supply system.
PLoS ONE 17(3): e0264079. https://doi.org/10.1371/journal.pone.0264079. Accessed
21 October 2022.
Balisacan AM (1994) Demand for food in the Philippines: responses to price and income
sumchanges. Philippine Review of Economics and Business 31(2):137–163.
BAS [Bureau of Agricultural Statistics]. 2013. Estimating price and income elasticities of
demand of selected food commodities in the Philippines. Paper prepared by the Socio-
Economic Statistics Section, Agricultural Accounts and Statistical Indicators Division,
Bureau of Agricultural Statistics. Quezon City: PSA.
Bayudan-Dacuycuy, C., L. Baje, and M. Almario. 2019. ECCD-F1KD Situation Analysis in
Selected KOICA-UNICEF Municipalities in Northern Samar. Discussion Paper Series
No. 2019-29. Quezon City: PIDS.
Blundell, R, P. Pashardes, G. Weber. 1993. What do we learn about consumer demand patterns
from micro data. American Economic Review 83:570-597.
Bouis, H. 1982. Demand for Cereal Staplesin the Philippines. Paper presented at the Rice
Policiesin Southeast Asia Project Workshop, Jakarta, August 17-20.
Bouis, H. 1990. Evaluating demand for calories for urban and rural populations in the
Philippines: implications for nutrition policy under economic recovery. World
Development 18(2):281-299.
Bouis, H., L Haddad, E. Kennedy. 1992. Does it matter how we survey demand for food?
Evidence from Kenya and the Philippines. Food Policy 17(5):349 – 360.
Bouis, H, and L. Haddad. 1992. Are estimates of calorie-income elasticities too high? A
recalibration of the plausible range. Journal of Development Economics 39:333-364.
37
Canlas, D. 1983. Estimating Price and income Elasticities of the Demand for Food: The
Philippines,1965. Paper presented during the Nutrition Symposium-Seminar Series.
Quezon City: University of the Philippines School of Economics, July 6.
Capuno, J. and A. Orbeta, eds. Human Capital and Development in the Philippines: a
Festschrift in Honor of Alejandro N. Herrin. PIDS Book 2008-01. Quezon City: PIDS.
Cho, Y, and D Johnson. 2022. COVID-19 and Social Assistance in the Philippines: Lessons
for Future Resilience Yoonyoung Cho and Doug Johnson. Philippine Country Office.
Washington, D.C: World Bank.
Deaton, A, and J. Muellbauer. 1980. An almost ideal demand system. American Economic
Review 70:312-326.
DOST-DOST-FNRI [Department of Science and Technology - Food and Nutrition Research
Institute]. 2022a. 2018-2019 Expanded National Nutrition Survey (ENNS) Food
Consumption Survey. Philippine Nutrition Facts and Figures. Taguig City: DOST-
DOST-FNRI.
DOST-DOST-FNRI. 2022b. 2018-2019 Expanded National Nutrition Survey (ENNS).
Philippine Nutrition Facts and Figures. Taguig City: DOST-DOST-FNRI.
DWSD and WFP. 2022. Joint Assessment of the Impact of COVID-19 Pandemic on Food
Security and Essential Needs and the Role of the Social Amelioration Program in the
Philippines. Quezon City: DSWD.
Fujii, T., 2016. Is urban food demand in the Philippines different from China? Chinese global
production networks in ASEAN. Research Collection School Of Economics 2-2016,
97-118.
Garcia, Y, M Dey, S. Navarez. 2005. Demand for fish in the Philippines: a disaggregated
analysis. Aquaculture Economics and Management 9(1-2):141-168.
Goldman, H, and C. Ranade. 1976. Food Consumption Behavior by Income Class in Rural and
Urban Philippines. Occasional Paper 90. Ithaca: Department of Agricultural
Economics, Cornell University.
Greene, W. 2017. Econometric Analysis. 8th ed. London: Pearson.
Heien, D., and C Wessels. 1990. Demand systems estimation with microdata: A censored
regression approach. Journal of Business and Economic Statistics 8(3):365-371.
Herrin, A. 1992. Micro impacts of macroeconomic adjustment policies on health, nutrition and
education. Working Paper Series No. 92-18. Quezon City: PIDS.
Herrin, A. 2016. Putting Prevention of Childhood Stunting into the Forefront of the Nutrition
Agenda: A Nutrition Sector Review. Discussion Paper Series No. 2016-21. Quezon
City: PIDS.
Herrin, A., M. Abrigo, Z. Tam, and D. Ortiz. 2018. Child Stunting Prevention: The Challenge
of Mobilizing Local Governments for National Impact. Discussion Paper Series No.
45. Quezon City: PIDS.
Huang, J, and C David. 1993. Demand for cereal grains in Asia: the effect of urbanization.
Agricultural Economics 8:107-124.
Lantican, F, M. Sombilla, K. Quilloy. 2013. Estimating the demand elasticities of rice in the
Philippines. Monographs. Los Baños, Laguna: SEARCA [Southeast Asian Regional
Centre for Graduate Study and Research in Agriculture.]
38
Malabayabas, ML, Yorobe, J, and N. De Castro. 2009. Household demand analysis for fresh
meat in the Philippines. Philippine Journal of Veterinary and Animal Sciences
35(2):168-176.
Mutuc, M.E, S. Pan, R. Rejesus. 2007. Household vegetable demand in the Philippines: is there
an urban-rural divide? Agribusiness 23:511-527.
Orbeta, A., and M. Alba. 1998. Simulating the impact of macroeconomic policy changes on
macronutrient availability in households. Journal of Philippine Development 25(1):
181-227.
Pante, F. Consumer Demand Functions: An Empiricial Evaluation of Alternative Function
Forms." Ph.D. dissertation. Quezon City: University of the Philippines School of
Economics.
Pineda, V., and J. Berido. 1998. The Case of Butuan City: A Consistent Regional Outstanding
Winner in Nutrition (CROWN). Discussion Paper Series No. 1998-33. Quezon City:
PIDS.
Pitt, N. 1983. Food preferences and nutrition in rural Bangladesh. Review of Economics and
Statistics 64(1):105-114.
PSA. 2021. Average Annual Per Capita Consumption of Selected Agricultural Commodities.
Quezon City: PSA.
PSA. 2022. PSA Openstat. http://openstat.psa.gov.ph.
Quisumbing, A, T. Valerio, E. Red, G. Villavieja. 1988. Flexible functional form estimates of
Philippine demand elasticities for nutrition policy simulation. Working Paper Series
No. 88-13. Quezon City: PIDS.
Ulep, V., J. Uy, and L. Casas. 2021. What explains the large disparity in child stunting in the
Philippines? A decomposition analysis. Public Health Nutrition 1 - 36
World Health Organization (WHO) (2018) Increasing Fruit and Vegetable Consumption to
Reduce the Risk of Non-Communicable Diseases. World Health Organization, E-
Library of Evidence for Nutrition Actions (eLENA) World Health Organization,
Geneva, Switzerland.
39
10. Appendix Tables
Appendix Table 1: Parameters of the Demand Model
Explanatory Dependent variable (in expenditure shares)

variables Rice Other cereals Fish Meat Poultry Fruit Vegetables Dairy Fats & oils
Rice 0.1052 0.0245 -0.0331 -0.0172 -0.0121 0.0062 -0.0218 -0.0001* -0.0057
Other cereals 0.0245 -0.0494 -0.0078 0.0002* 0.0072 0.0001 -0.0035 -0.0002* 0.0004*
Fish -0.0331 -0.0078 0.1006 -0.0022 *
-0.0029 0.0065 0.0028 -0.0001* -0.0020
Meat -0.0172 0.0002* -0.0022 0.0487 -0.0148 0.0021 0.0104 -0.0036 -0.0023
Poultry -0.0121 0.0072 -0.0029 -0.0148 0.0187 -0.0001* 0.0082 -0.0045 -0.0036
Fruit 0.0062 0.0001 0.0065 0.0021 -0.0001 *
-0.0518 0.0024 -0.0023 0.0020
Vegetables -0.0218 -0.0035 0.0028 0.0104 0.0082 0.0024 0.0124 0.0017 -0.0032
Dairy -0.0001* -0.0002* -0.0001 -0.0036 -0.0045 -0.0023 0.0017 -0.0028* -0.0010*
Fats & oils -0.0057 0.0004 -0.0020 -0.0023 -0.0036 0.0020 -0.0032 -0.0010* 0.0133
Food expenditure 0.0160 0.0729 0.0244 0.0428 0.0307 0.0262 0.0425 0.0206 0.0248
Quadratic term 0.0039 0.0083 0.0082 0.0083 0.0005 0.0005* 0.0008* 0.0059 0.0030
IMR -0.1401 0.0174 -0.2248 -0.0424 0.0427 0.0707 -0.1653 0.0182 0.0117
4Ps 0.0221 0.0032 -0.0112 -0.0081 -0.0038* -0.0001* 0.0000* -0.0062 0.0008*
GMP 0.0046* 0.0006 -0.0032* -0.0058* -0.0153 0.0019* -0.0014* 0.0281 -0.0039
Age -0.0006 -0.0001* 0.0010 -0.0018 -0.0033 0.0002 0.0002 -0.0001 0.0000
Education -0.0199 0.0000 -0.0013 0.0053 0.0077 0.0027 -0.0001* 0.0041 0.0015
Marital status 0.0270 0.0006 0.0012 *
-0.0150 -0.0124 -0.0035 0.0027* -0.0066 0.0009
Occupation -0.0095 -0.0043 0.0023* 0.0097 0.0069 0.0010* -0.0057 0.0057 0.0008
Urbanity -0.0097 0.0020 -0.0082 0.0094 0.0077 0.0013* -0.0083 -0.0001* -0.0010*
Wealth -0.0466 0.0033 -0.0070 0.0349 0.0158 0.0032 -0.0099 0.0124 0.0019
"R-squared" 0.37 0.16 0.09 0.13 0.06 0.03 0.06 0.05 0.13
40

Food and Nutrient Intake Response To Food Prices and Govt Programs

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Food and Nutrient Intake Response To Food Prices and Govt Programs

Uploaded by

Copyright:

Available Formats

DECEMBER 2022

DISCUSSION PAPER SERIES NO. 2022-44

Food and Nutrient Intake Response to Food

18th Floor, Three Cyberpod Centris - North Tower publications@pids.gov.ph

PHILIPPINE INSTITUTE FOR DEVELOPMENT STUDIES

2. Recent economic trends ............................................................................................... 2

3. Recent trends in food consumption and nutrition ..................................................... 5

4. Review of literature and methodology ....................................................................... 17

5. Empirical strategy for household demand ................................................................ 20

6. Results of demand estimation .................................................................................... 26

7. Simulated impacts of recent economic shocks: household level .......................... 32

10. Appendix Tables ...................................................................................................... 40

2. Recent economic trends

2.2 Output and employment

Source: PSA (2022).

The Bayanihan programs of government provided substantial social protection to households in

Table 2: Annual inflation rate, 2019 – 2022 (%)

Source: PSA (2022).

3. Recent trends in food consumption and nutrition

3.1 Food consumption and expenditure

Table 4: Per capita consumption of food items, 1978 – 2019

Food consumption is concentrated in animal proteins and cereals; increasing wealth is

Sources: DOST-FNRI (2022a).

3.2 Nutritional status

Sources: DOST-FNRI (2022).

Source: DOST-FNRI (2022b).

Source: DOST-FNRI (2022b).

3.3 Nutrient intake

Source: Mbuya et al (2021).

Source: Mbuya et al (2021).

Table 9: Contribution to energy and nutrient intake of households, by food group,

Age group is a key determinant of energy and nutrient intake adequacy.

3.4 Government programs

4. Review of literature and methodology

4.1 Conceptual framework

Consumption is motivated by maximization of a utility function U h (q1h , q2 h ,..., qMh ) . Under

An estimate of point-response to a change in consumer price is the elasticity of consumption

4.2 Studies focused on nutrient intake

4.3 Consumption studies

4.4 Role of government programs

5. Empirical strategy for household demand

5.1 Data source

5.2 Case of purchased food groups

index ( Ph = ∑ s jh ln p jh ); and zkh denotes other control variables corresponding to household

demographics, location, and survey period. Parameters to be estimated are α i , γ ij , βi , λi , δ i ,

Case of zero purchases for some food groups

µ0i + ∑ µil xlh +ν ih , ν ih ~ N (0,1)

Incorporating food price and expenditure in the selection model

Suppose (5) is specified as follows (non-price and expenditure terms suppressed):

Simulations of nutrient intake and individual demand

Denote “%” as the percentage operator, e.g. dq q = %q ; this implies:

%qi | dmh = %mh (η mih )

6. Results of demand estimation

6.1 The selection model for household food demand

Table 15: Coefficients of price and expenditure variables, probit regressions

Table 16: Coefficients of price and expenditure variables, probit regressions

6.2 Food demand estimation

Parameters of the household food demand system

Income elasticities of household food demand

Price elasticities of household food demand

Table 18: Price elasticities of household demand, with selection effects

Table 19: Price elasticities of household demand, without selection effects